Oligopeptide transporter PepT1 in Atlantic cod (Gadus morhuaL.): cloning, tissue expression and comparative aspects

Uptake of oligopeptides (di- and tripeptides) into cells is due to membrane transport proteins that belong to the so-called SoLute Carrier 15 (SLC15)family (Daniel and Kottra,2004). One of the members of this family, namely PepT1 (or SLC15A1), is highly expressed in the intestine of vertebrates, where it is responsible for the transport of a significant fraction of dietary protein across the brush-border membrane of the small intestinal epithelium (for a review, see Daniel, 2004). PepT1 functions as a Na⁺-independent, H⁺-dependent,H⁺-coupled transporter of a variety of di- and tripeptides. PepT1 is also responsible for the transport of orally active drugs, such asβ-lactam antibiotics, aminopeptidase and angiotensin-converting enzyme inhibitors, δ-aminolevulinic acid, and many selected pro-drugs (for a review, see Rubio-Aliaga and Daniel,2002).

PepT1 proteins have been characterized in great detail in higher vertebrates [mostly in mammals, but also in birds(Daniel et al., 2006)]. In contrast, information about these proteins in lower vertebrates is very limited, with the sole exception of the PepT1-type oligopeptide transporter of the teleost zebrafish Danio rerio, whose functional activity and pattern of expression in tissue has been assessed(Verri et al., 2003). In particular, zebrafish PepT1 is highly expressed in the proximal intestine,where it mediates the uptake of a large amount of di- and tripeptides derived from the protein digestion process.

We report here the cloning of a full-length cDNA that encodes for the PepT1-type oligopeptide transporter of the Atlantic cod (Gadus morhuaL.). This is an important commercial species in many North Atlantic countries,and has recently been targeted for aquaculture, mainly due to depletion of natural stocks by overfishing (e.g. Brander, 2007). For this reason, Atlantic cod has become an important model fish species. The complete Atlantic cod PepT1 sequence, the second fish sequence resolved to date, has made a detailed comparison along the vertebrate series possible, allowing identification of highly conserved motifs/regions in all the vertebrate PepT1 transporters. It also became possible to study tissue expression as well as the regional distribution in the digestive tract of the transporter at the mRNA level.

Atlantic cods (Gadus morhua L.) used in this experimental work (45 cm in length) were reared in commercial systems, or at the Bergen High-Technology Centre (Bergen, Norway). Fish were routinely fed to satiety once daily, and fasted for 36–48 h prior to sampling. The fish were killed by a blow to the head and their organs were rapidly removed,transferred into RNAlater (Ambion, Austin, TX, USA) and stored at–80°C until RNA extraction. In two fish, the gut was dissected into ten sections for regional analysis of the spatial distribution along the intestinal tract. The segments were stomach, pyloric area, pyloric caeca(inner and outer segments), and six segments in the remainder of the intestine based on divisions of the three loops present in the dissected gut. The last segment (segment 10) also comprised the hindgut.

Total RNA was isolated from cod intestine using the acid guanidinium thiocyanate-phenol-chloroform method(Chomczynski and Sacchi, 1987). Total RNA (1 μg) was reverse transcribed at 37°C for 1 h using Moloney Murine Leukemia Virus Reverse Transcriptase (M-MLVRT; Invitrogen, Carlsbad,CA, USA) and a reverse oligo(dT)-adapter[5′-ACGCGTCGACCTCGAGATCGATG(T)₁₈-3′]. Amplification of a partial Atlantic cod PepT1 (codPepT1) was performed by polymerase chain reaction (PCR) using Taq DNA polymerase and degenerate primers whose design was based on conserved regions of known vertebrate PepT1 genes. Among these, the sequences from icefish (Chionodraco hamatus; GenBank accession no. AY170828), European eel (Anguilla anguilla; GenBank accession no. AY167576) and zebrafish (Danio rerio; GenBank accession no. AY300011). In addition, by data-mining EST sequences and viamanual and automated(http://genes.mit.edu/genomescan.html)predictions of locations and exon–intron boundaries, respectively, the partial cDNA for Salmo salar (TPA Database accession no. BK004882)and for Takifugu rubripes (TPA Database accession no. BK004883) were outlined and included in the design of the following degenerate primers:forward primer-1 (FP-1; 5′-GCDGCMTTYGGDGGAGAYCAGTT-3′) and reverse primer-1 (RP-1; 5′-CCAGTCCAKCCAGTGCKCYCTYTTKKGG-3′). The PCR reactions were conducted for 30 cycles (1 cycle: 30 s at 95°C, 45 s at 58°C and 1 min at 72°C), followed by a 10 min final extension at 72°C, with Taq DNA polymerase (Invitrogen). The PCR products were size-fractionated in a 2% agarose gel by electrophoresis and the fragment of the expected size was gel-purified using the GFX PCR DNA and Gel Band Purification Kit (Amersham Biosciences, Uppsala, Sweden), ligated into the pGEM-T Easy vector (Promega, Madison, WI, USA), and multiple clones were screened by sequencing. The identity of the resulting PCR product was confirmed by DNA sequence analysis at Macrogen Inc. (South Korea).

The full-length sequence of codPepT1 was obtained by 3′ and 5′RACE using the Marathon cDNA Amplification Kit (Clontech, Mountain View, CA,USA) with Advantage Klen Taq polymerase (Clontech). PolyA(+) RNA was purified with the Oligotex mRNA Midi Kit (Qiagen GmbH, Hilden, Germany). 1μg of this mRNA was used to construct one Marathon cDNA library (Clontech)and the 5′ and 3′ ends were obtained by rapid amplification of cDNA ends (RACE)-PCR using Advantage Klen Taq polymerase, the AP1 primer and gene-specific primers codPepT1-F1(5′-CTCCATCTTCTACCTGTCCATCAACGCA-3′) and codPepT1-R1(5′-GCTACGGTTCCTGAAGCGGTTTTTGACT-3′). Amplification conditions were those suggested by the supplier. PCR products were size-separated by agarose gel electrophoresis and visualized by ethidium bromide staining,extracted from the gel with GFX PCR DNA and the Gel Band Purification Kit, and cloned into the pGEM-T Easy vector. Final identification was made by DNA sequence analysis at Macrogen Inc.

The codPepT1 amino acid sequence was deduced using the open reading frame(ORF) finder program at http://www.ncbi.nlm.nih.gov. Putative transmembrane domains were predicted using TMHMM 2.0(http://www.cbs.dtu.dk/services/TMHMM-2.0/),which is part of the Simple Modular Architecture Research Tool (SMART; at http://www.expasy.org/prosite/). Potential N-glycosylation, cAMP/cGMP-dependent protein kinase and protein kinase C recognition sequences were identified using the PROSITE 19.7 computational tools(http://www.expasy.org/prosite/).

Nucleotide sequences were routinely compared with the GenBank database using the BLAST algorithm (Altschul et al.,1997). Clustal W 1.82 was used to align amino acid sequences(www.ebi.ac.uk/clustalw). The phylogenetic reconstruction was generated using the neighbor-joining (NJ)method (Saitou and Nei, 1987),as implemented in MEGA 3.0 (Kumar et al.,2004). Phylogenetic trees were constructed with bootstrap confidence values based on 1000 replicates. GenBank accession numbers for amino acid sequence comparisons are: AAQ65244 [zebrafish PEPT1(Verri et al., 2003)],AAA17721 [rabbit PEPT1 (Fei et al.,1994)], NP_001003036 (dog PEPT1), AAO43094 [pig PEPT1(Klang et al., 2005)],XP_599441 (bovine PEPT1), AAB61693 [human PEPT1(Liang et al., 1995)],NP_001028071 [macaque PEPT1 (Zhang et al.,2004a)], BAA09318 [rat PEPT1(Miyamoto et al., 1996)],NP_444309 [mouse PEPT1 (Fei et al.,2000)], AAK14788 [sheep PEPT1(Pan et al., 2001)], AAK39954[chicken PEPT1 (Chen et al.,2002)], AAO16604 (turkey PEPT1).

Total RNA was purified from heart, spleen, gill, eye, intestine, ovary,kidney and liver using the acid guanidinium thiocyanate-phenol-chloroform method (Chomczynski and Sacchi,1987), and subjected to DNase treatment (Turbo DNase, Ambion). For regional analysis of the intestinal tract, total RNA was extracted from various gut segments, as depicted in Fig. 5 (see also Animals and tissue sampling). cDNA was synthesized starting from 1.5 μg total RNA, using SuperScript III(Invitrogen) and following the manufacturer's recommen dations. The following Atlantic cod specific primers were designed on the basis of the nucleotide sequence GenBank accession no. CO541820 for Gadus morhua elongation factor 1α (EF1α) mRNA: forward primer CodEF1AF(5′-CCCCTCCAGGACGTCTACAAG-3′) and reverse primer CodEF1AR(5′-GGCAGAGCCACCGATCTTC-3′), while the forward primer CodPepT1F(5′-CCGCTTCAGGAACCGTAGC-3′) and the reverse primer CodPepT1R(5′-TTCGCTGTCATATCTTCGTACGA-3′) were used for amplification of Atlantic cod PepT1. The PCR reactions were conducted for 35 cycles (1 cycle:45 s at 95°C, 30 s at 60°C and 50 s at 72°C), followed by a 10 min final extension at 72°C, with Taq DNA polymerase. The PCR products were size-fractionated by agarose gel electrophoresis.

Tracts of cod intestine were dissected out, cut into smaller pieces, and fixed in 4% paraformaldehyde in 0.1 mol l^–1 phosphate buffer(PB: 0.028 mol l^–1 NaH₂PO₄, 0.071 mol l^-1 Na₂HPO₄, pH 7.2) for 16 h at 4°C. Buffer was changed twice to 25% sucrose in PB for 30 min and 15 min, and samples were kept in 25% sucrose in PB with 10% Tissue Tek (Miles Inc.,Elkhart, IN, USA) overnight at 4°C. The tissues were embedded in Tissue Tek and stored at –80°C. Cryostat sections (10 μm thick) were collected on frozen Superfrost Plus glass slides (Merck, Darmstadt, Germany),air-dried and stored at –20°C until use.

The nucleotide fragment covering amino acids 254–480 of Atlantic cod PepT1 sequence was amplified by PCR using the forward primer CodPepT1F(starting at nucleotide 824) and the reverse primer CodPepT1R (starting at nucleotide 1504; see Fig. 1). The amplification product was subcloned into the pCR4-TOPO vector (Invitrogen)and the recombinant clone sequenced to confirm the identity of the insert. In situ analysis was performed with digoxigenin (DIG)-labeled cRNA probes (sense and antisense), using a DIG RNA labeling Kit (Roche, Mannheim,Germany). The DIG-labelled antisense riboprobe was synthesized in vitro with T7 RNA polymerase using the SpeI-cleaved recombinant pCR4-TOPO clone. The corresponding sense riboprobe was synthesized with T3 RNA polymerase using the NotI-cleaved recombinant pCR4-TOPO clone. DIG incorporation and the concentration of the probes were analyzed by spot tests(Roche).

In situ hybridization procedures on tissue sections were carried out according to published methods(Ebbesson et al., 2005).

The Atlantic cod PepT1 cDNA (GenBank accession no. AY921634) was 2838 bp long, with an open reading frame of 2190 bp encoding a putative protein of 729 amino acids (Fig. 1). Hydropathy analysis predicted at least 12 potential membrane-spanning domains with a large extracellular loop between transmembrane domains IX and X(Fig. 1). Five putative extracellular N-glycosylation sites (Asn¹²⁴, Asn⁴²²,Asn⁴⁴⁶, Asn⁴⁵⁶ and Asn⁴⁹⁹) and three putative intracellular cAMP/cGMP-dependent protein kinase phosphorylation sites(Thr³⁶⁹, Ser⁶⁹¹ and Ser⁷⁰⁰) were identified(Fig. 1). No putative protein kinase C phosphorylation sites were found. When compared with PepT1-type members of the SLC15 family already characterized from other vertebrates(Fig. 2), the predicted Atlantic cod PepT1 amino acid sequence exhibited a higher percentage of identity with the zebrafish PepT1 (63%) than with any other PepT1-type transporter characterized so far in higher vertebrates (58–61%). The phylogenetic reconstruction of vertebrate PepT1-type proteins clustered codPepT1 to the `fish' branch of the phylogenetic tree (together with zebrafish PepT1) and indicated early divergence both of the Atlantic cod sequence with respect to that of the zebrafish and of the two fish protein sequences with respect to those of the tetrapod group(Fig. 3).

To date, codPepT1 is the second full-length cDNA encoding for a PepT1-type transporter that has been cloned from a teleost fish and made available via public databases. Like zebrafish PepT1(Verri et al., 2003), it is evolutionarily distant from the sequences already characterized from higher vertebrates, i.e. birds and mammals (for a review, see Daniel et al., 2006), and thus represents a novel reference sequence to support comparative analysis along the vertebrate series. With this in mind, alignment and comparison of the deduced amino acid sequence of Atlantic cod PepT1 to those of the other PepT1-type members of the SLC15 family already characterized in vertebrates were performed in order to recognize and recapitulate major structural/functional features common to vertebrate PepT1-type proteins(Fig. 2).

The first examination included analysis of putative membrane-spanning domains, N-glycosylation sites, cAMP/cGMP-dependent protein kinase phosphorylation sites and protein kinase C phosphorylation sites. Examination of the predicted membrane-spanning domains revealed 11 (out of 12) highly conserved regions of hydropathy among the proteins (membrane-spanning domains II–XII). Only in one case, i.e. transmembrane domain I, did the predicted membrane-spanning regions of the different proteins not (fully)overlap, but resulted in either of two adjacent sub-regions of hydropathic amino acids (Fig. 2; compare chicken, turkey, sheep, dog, human, rat, mouse and rabbit to macaque,zebrafish and Atlantic cod PepT1 sequences) or in the two sub-regions of hydropathic amino acids (Fig. 2; see bovine and pig PepT1 sequences). Such sub-regions were indicated as membrane-spanning domain Ia and membrane-spanning domain Ib, and may be indicative of a larger than expected area of hydrophobic amino acids in close proximity to the N terminus of the PepT1-type proteins. Analysis of single N-glycosylation, cAMP/cGMP-dependent protein kinase phosphorylation and protein kinase C phosphorylation predicted sites revealed (1) a large N-glycosylation-rich region within the large extracellular loop between membrane-spanning domains IX and X; (2) a single highly conserved cAMP/cGMP-dependent protein kinase phosphorylation motif close to the putative spanning domain IX (see Thr³⁶⁹ in the Atlantic cod sequence as a reference) that is present in every PepT1 sequence except the human (in which Ala³⁵⁷ is present instead of threonine); and (3) an adjacent protein kinase C phosphorylation (Ser-Leu-Lys) motif that is present in mammalian but not in avian and fish sequences. In the end, comparisons of tetrapod and fish sequences revealed at least one interesting conserved and previously undescribed stretch (of 8–12 amino acid residues) within the large extracellular loop, roughly centered around amino acid residues 480–491 in the Atlantic cod sequence. Further studies are necessary to assess the structural/functional significance of this stretch. Other regions/residues that are recognized as structurally or functionally important within the mammalian PepT1 primary structure (for example, see Daniel, 2004) were invariably conserved in Atlantic cod PepT1, as well as the PepT1 sequence of the zebrafish. These include the `PTR2 family proton/oligopeptide symporters signature 1' motif (PROSITE pattern: PS01022; amino acid residues 77–101 in Atlantic cod PepT1) and `PTR2 family proton/oligopeptide symporters signature 2' motif (PROSITE pattern: PS01023; amino acid residues 170–182 in Atlantic cod PepT1).

The next examination involved comparative analysis to confirm identity along the vertebrate series of those amino acids for which mutational studies in mammalian systems had previously assessed a significant effect on function. Amino acids along the sequences for which mutational analysis was performed are highlighted in Fig. 2. All of them fall within putative membrane-spanning domains (namely membrane-spanning domains I, II, IV, V, VII, VIII and X) and are invariably conserved (identities or conservative substitutions) among the species. A list of the effects of a single amino acid substitution, as obtained by site-directed mutagenesis, on the function of human, rabbit and rat PepT1 transporters, is reported in Table 1.

Expression of Atlantic cod PepT1 was analysed in various tissues/organs of adult fish. Using Atlantic cod PepT1-specific primers, a 681 bp RT-PCR product was amplified from total RNA isolated from intestine, kidney and spleen. No signal was obtained from samples of heart, gill, eye or liver, while a slight signal was obtained from ovary. As a control to assess RNA quality, EF1αRNA amplification was performed using Atlantic cod EF1α-specific primers, which invariably gave comparable 603 bp amplification products for all tissues tested (Fig. 4).

Spatial distribution of Atlantic cod PepT1 expression in the digestive tract was studied in detail (Fig. 5). A 681 bp RT-PCR product was amplified from total RNA isolated from a total of ten segments (see Fig. 5A) along the intestine of adult fish. No signal was obtained from the stomach (segment 1; see Fig. 5B), while strong signals were obtained in the following segments. There appeared to be a weakening in the signal from segment 9 (the last 1/5 of the intestine), but in the last segment, which included the hindgut, the expression was very low (see Fig. 5B). The control for RNA quality (EF1α RNA amplification)invariably gave comparable 603 bp amplification products for all tested segments.

Expression of PepT1 mRNA at the intestinal level was further analyzed by in situ hybridization (Fig. 6). Using the DIG-labeled antisense probe, expression of the PepT1 mRNA was detected only in the epithelial layer of the intestine of Atlantic cod (see Fig. 6B), while the sense probe revealed no staining (see Fig. 6A).

We have identified a full-length cDNA from Atlantic cod that encodes for a novel PepT1-type peptide transporter (Fig. 1). The predicted protein, designated as codPepT1, shares a high overall identity with PepT1-type transporters (58–63%) when compared to other known PepT1 proteins in vertebrates(Fig. 2), and clusters to the`fish' branch of the reconstructed phylogenetic tree, together with zebrafish PepT1 (Fig. 3). To date,codPepT1 represents the second PepT1-type peptide transporter recognized from a teleost fish [after zebrafish PepT1(Verri et al., 2003)], thus representing an additional molecular tool for the understanding of evolutionary and functional relationships among vertebrate peptide transporters.

In this paper, the comparative analysis of two fish (Atlantic cod and zebrafish) amino acid sequences vs a collection of sequences from various tetrapod (two avian and nine mammalian) species has allowed the recognition of a number of highly conserved motifs, regions and protein domains (and the exclusion of others) simply on the basis of the overall sequence alignment (see Fig. 2), as well as a better representation of some still unresolved areas of the protein. It is noteworthy that the individuation of two adjacent sub-regions of hydropathic amino acids instead of one single membrane-spanning domain I is suggestive of the existence of a very large hydrophobic area close to the N terminus of the PepT1 proteins. Moreover, the occurrence, beside the large N-glycosylation-rich region, of short, well-conserved stretches of amino acids within the big extracellular loop is novel structural evidence that deserves further study. The only highly conserved cAMP/cGMP-dependent protein kinase phosphorylation motif close to membrane-spanning domain IX (found in all vertebrate sequences with the exception of the human) and the adjacent protein kinase C phosphorylation motif (found in mammalian sequences only)would also merit further analysis, mostly in the light of the well-documented regulation of PepT1 protein activity in Caco-2 cells, mammalian models and human intestine, via agonists or antagonists of protein kinase A and C and hormones/extracellular signals (for a review, see Daniel, 2004). It is also noteworthy that none of the PepT1 protein kinase phosphorylation motifs have been analyzed so far with respect to their ability to functionally transmit the observed effects of second messengers at the protein level.

Functional analysis of mammalian PepT1 transporters has led to a general scheme of peptide transport and to the concept that such transport occurs in virtually all vertebrates according to the basic design (in terms of mode of transport, kinetics, proton- and membrane potential-dependence, pH dependence,electrogenicity, substrate specificity, protein sorting to the membrane, etc.)defined in the mammalian systems (for reviews, see Brandsch et al., 2004; Daniel, 2004; Daniel and Kottra, 2004; Daniel et al., 2006; Terada and Inui, 2004). Interestingly, the structural/functional characterization of zebrafish PepT1 partly contradicted this assumption, revealing that this transporter closely resembles mammalian systems in terms of low-affinity/high-capacity properties of the transport, but also exhibits some peculiarities, such as a unique pH dependence (Verri et al.,2003). It also suggested that fish PepT1 proteins might have allowed a useful comparative approach to focus on amino acid residues, motifs,conserved regions and protein domains that are relevant to the general function of the transporter or, conversely, to the determination of a species-specific phenotype. Since 1997, the function of mammalian PepT1 proteins has been progressively investigated with the support of single amino acid mutational analysis to define relevant structural/functional motifs or domains involved in the manifestation of the phenotype (see Table 1 and references therein). Such investigation allowed the identification of many amino acids that play a crucial role in the function of the transporter, although it was performed on an almost random basis and focused primarily on the functional role of selected transmembrane domains. In this respect, using our comparative analysis, we confirmed identity or conservative substitution along the vertebrate series of all those amino acids (23 amino acid residues distributed in membrane-spanning domains I, II, IV, V, VII, VIII and X; see Fig. 2) for which functional mutational analysis had established a significant effect on function. In perspective, such a comparative approach might help to rationalize the selection of the most suitable amino acid residues to target in site-directed mutagenesis experiments.

Atlantic cod PepT1 is highly expressed in the intestine, and a significant RT-PCR signal was also found in both kidney and spleen tissue(Fig. 4). Interestingly, this is the same pattern of expression as is found in zebrafish for PepT1(Verri et al., 2003), which confirms a common theme among fish. A more detailed analysis of the regional distribution along the intestinal tract of cod, as performed herein for the first time in an adult fish to our knowledge, revealed that PepT1 is ubiquitously expressed in all segments after the stomach, including the pyloric caeca (Figs 5 and 6). This suggests that Atlantic cod may have a very high capacity to absorb small peptides from dietary protein digestion, with absorption occurring in most parts of the intestine. The low expression in the last segment that included the hindgut indicates that this segment is not, or only slightly, involved in peptide absorption. However, it may be involved in final adjustments of ion and water composition(for a review, see Marshall and Grosell,2005).

In conclusion, the Atlantic cod intestinal PepT1-type oligopeptide transporter has been identified and characterized with respect to its expression in tissues. On the basis of the overall amino acid sequence alignment, conserved amino acids and novel motifs, regions and protein domains have been recognized in all vertebrates. In a perspective, Atlantic cod PepT1 can represent a useful tool for the study of gut regionalization, as well as a marker for the functional analysis of temporal and spatial expression during ontogeny, under the effects of various dietary sources, and in pathological states.

Supported by a study leave grant from the University of Bergen to I.R. and Research Council of Norway grants 165203/S40, 174229/S15 and 175021. A.R. and T.V. acknowledge support from grants from the University of Salento (Fondi ex–60%, 2006) and from the Apulian Region (Progetto Strategico, cod. Cip PS_070, and Progetto Esplorativo, cod. Cip PE_062). Thanks to Dr V. Laize(UALG-CCMAR), V. Tronci and Dr C. Jolly (UoB) for additional technical assistance. I.R. thanks M. Leonor Cancela and her group for creating excellent working conditions during his sabbatical stay at CCMAR.